iascl_beh_analysis
Jojo Hu
10/10/2019
IASCL Behavioral Analysis
Age at MRI and EEG
#Clean Data
All behavioral data prep
MRI correlation Data Prep
MRI Subjects, TSL LSL only
mri <- read.csv("/Users/jojohu/Documents/Qlab/iascl/RH_LH_GM_extraction.csv")
mri <- mri[,c(1,2,5,6)]
mri_gray_matter_wide <- cast(mri, part_id~roi, value = "total_gray_matter_volume")
mri_gray_matter_wide <- mri_gray_matter_wide[,!colnames(mri_gray_matter_wide) %in% "LeftAntTemp"]
average_cor_thick_wide <- cast(mri, part_id~roi, value = "average_cortical_thickness")
average_cor_thick_wide <- average_cor_thick_wide[,!colnames(average_cor_thick_wide) %in% "LeftAntTemp"]
iascl_beh_gray_matter <-
merge(mri_gray_matter_wide, blast_spoli_data_wide,
by.x = "part_id", by.y = "part_id",
all.x = TRUE)
iascl_beh_gray_matter <-
merge(iascl_beh_gray_matter, gjt[,c("part_id", "a_prime")],
by.x = "part_id", by.y = "part_id",
all.x = TRUE)
iascl_beh_gray_matter <-
iascl_beh_gray_matter[,c(1:12, 25, 36, 16, 17, 22, 23, 27, 28, 31, 32)]
iascl_beh_gray_matter <-
merge(iascl_beh_gray_matter, drive_checklist[,c("Participant.ID", "age_at_neuro_year")],
by.x = "part_id", by.y = "Participant.ID",
all.x = TRUE)
iascl_beh_gray_matter_td <-
iascl_beh_gray_matter[which(iascl_beh_gray_matter$group == "TD"),]
iascl_beh_gray_matter_asd <-
iascl_beh_gray_matter[which(iascl_beh_gray_matter$group == "ASD"),]
iascl_beh_cor_thick <-
merge(average_cor_thick_wide, blast_spoli_data_wide,
by.x = "part_id", by.y = "part_id",
all.x = TRUE)
iascl_beh_cor_thick <-
merge(iascl_beh_cor_thick, gjt[,c("part_id", "a_prime")],
by.x = "part_id", by.y = "part_id",
all.x = TRUE)
iascl_beh_cor_thick <-
iascl_beh_cor_thick[,c(1:12, 25, 36, 16, 17, 22, 23, 27, 28, 31, 32)]
iascl_beh_cor_thick <-
merge(iascl_beh_cor_thick, drive_checklist[,c("Participant.ID", "age_at_neuro_year")],
by.x = "part_id", by.y = "Participant.ID",
all.x = TRUE)
iascl_beh_cor_thick_td <-
iascl_beh_cor_thick[which(iascl_beh_cor_thick$group == "TD"),]
iascl_beh_cor_thick_asd <-
iascl_beh_cor_thick[which(iascl_beh_cor_thick$group == "ASD"),]
#MRI data participants
print(iascl_beh_cor_thick[,c("part_id", "group")])## part_id group
## 1 blast_c_019 TD
## 2 blast_c_059 TD
## 3 blast_c_061 ASD
## 4 blast_c_071 ASD
## 5 blast_c_095 ASD
## 6 blast_c_126 TD
## 7 blast_c_147 ASD
## 8 blast_c_165 TD
## 9 blast_c_168 TD
## 10 blast_c_206 TD
## 11 blast_c_207 TD
## 12 blast_c_210 TD
## 13 blast_c_213 TD
## 14 blast_c_215 <NA>
## 15 blast_c_216 TD
## 16 blast_c_224 TD#Intersection between Beh and MRI participants
intersect(blast_spoli_data_wide$part_id, iascl_beh_cor_thick$part_id)## [1] "blast_c_019" "blast_c_059" "blast_c_061" "blast_c_071" "blast_c_095"
## [6] "blast_c_126" "blast_c_147" "blast_c_165" "blast_c_168" "blast_c_206"
## [11] "blast_c_207" "blast_c_210" "blast_c_213" "blast_c_216" "blast_c_224"identical(iascl_beh_cor_thick$part_id, iascl_beh_gray_matter$part_id)## [1] TRUEEEG + MRI participants, SSL, TSL only
Task, group two-way Ancova controlling for age
#Extract participants that have all SL task completed
iascl_all_completed <-
eeg_mri_beh_data[,c(1:7,11:12,17:18,20,35,22,23,26,27,37:38)]
iascl_all_completed_acc <- iascl_all_completed[,c(1, 6, 2, 8, 9)]
iascl_all_completed_slope <- iascl_all_completed[,c(1, 6, 2, 16, 17)]
iascl_all_completed_acc<- melt(iascl_all_completed_acc,
id.vars = c("part_id", "group", "age_at_neuro_year"))## Warning: attributes are not identical across measure variables; they will
## be droppediascl_all_completed_slope<- melt(iascl_all_completed_slope,
id.vars = c("part_id", "group", "age_at_neuro_year"))## Warning: attributes are not identical across measure variables; they will
## be droppedcolnames(iascl_all_completed_acc)[colnames(iascl_all_completed_acc) == "variable"] <- "task"
colnames(iascl_all_completed_acc)[colnames(iascl_all_completed_acc) == "value"] <- "accuracy"
colnames(iascl_all_completed_slope)[colnames(iascl_all_completed_slope) == "variable"] <- "task"
colnames(iascl_all_completed_slope)[colnames(iascl_all_completed_slope) == "value"] <- "rt_slope"
#aov(accuracy~group*task + age+Error(subjid/task), data = data)
acc_task_group_aov <-
aov(accuracy~group*task + age_at_neuro_year + Error(part_id/task), data = iascl_all_completed_acc)## Warning in aov(accuracy ~ group * task + age_at_neuro_year + Error(part_id/
## task), : Error() model is singularsummary(acc_task_group_aov)##
## Error: part_id
## Df Sum Sq Mean Sq F value Pr(>F)
## group 1 0.0321 0.03205 2.169 0.1533
## task 1 0.0230 0.02301 1.557 0.2237
## age_at_neuro_year 1 0.0457 0.04565 3.089 0.0911 .
## group:task 1 0.0415 0.04153 2.810 0.1061
## Residuals 25 0.3695 0.01478
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Error: part_id:task
## Df Sum Sq Mean Sq F value Pr(>F)
## task 1 0.00195 0.001946 0.196 0.663
## group:task 1 0.00008 0.000082 0.008 0.929
## Residuals 18 0.17873 0.009929rt_slope_task_group_aov <-
aov(rt_slope~group*task + age_at_neuro_year + Error(part_id/task), data = iascl_all_completed_slope)## Warning in aov(rt_slope ~ group * task + age_at_neuro_year + Error(part_id/
## task), : Error() model is singularsummary(rt_slope_task_group_aov)##
## Error: part_id
## Df Sum Sq Mean Sq F value Pr(>F)
## group 1 0.001114 0.0011136 3.116 0.0898 .
## task 1 0.000036 0.0000363 0.102 0.7526
## age_at_neuro_year 1 0.000204 0.0002044 0.572 0.4566
## group:task 1 0.000001 0.0000007 0.002 0.9643
## Residuals 25 0.008936 0.0003574
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Error: part_id:task
## Df Sum Sq Mean Sq F value Pr(>F)
## task 1 0.000748 0.0007482 1.01 0.328
## group:task 1 0.000852 0.0008522 1.15 0.298
## Residuals 18 0.013333 0.0007407Gender Chi-square
iascl_all_completed_gender_wide <- iascl_all_completed[,c("part_id", "group", "sex")]
iascl_all_completed[,c(2:5, 8:19)] <- lapply(iascl_all_completed[,c(2:5, 8:19)], function(x) {
if (is.factor(x))
as.numeric(as.character(x))
else
x
})## Warning in FUN(X[[i]], ...): NAs introduced by coercioniascl_all_completed_gender_long <- melt(iascl_all_completed_gender_wide, id.vars = c("part_id", "group"))
gender_wide <- cast(iascl_all_completed_gender_long, group~value, length)
# Chi-square on gender
# p<0.05 tells us that they are not matched for gender between group
print(gender_wide)## group F M
## 1 ASD 1 11
## 2 TD 14 8print(chisq.test(gender_wide))##
## Pearson's Chi-squared test with Yates' continuity correction
##
## data: gender_wide
## X-squared = 7.5199, df = 1, p-value = 0.006102Gender is NOT matched between group.
Age difference between group
td_age <-
iascl_all_completed[which(iascl_all_completed$group == "TD"),
c("part_id", "group", "age_at_neuro_year")]
td_age$age_at_neuro_month <- td_age$age_at_neuro_year *12
asd_age <-
iascl_all_completed[which(iascl_all_completed$group == "ASD"),
c("part_id", "group", "age_at_neuro_year")]
asd_age$age_at_neuro_month <- asd_age$age_at_neuro_year *12
library("pastecs")##
## Attaching package: 'pastecs'## The following objects are masked from 'package:dplyr':
##
## first, last## The following objects are masked from 'package:data.table':
##
## first, lasttd_age_descrip_stat <- stat.desc(td_age)
td_age_descrip_stat <- td_age_descrip_stat[c("nbr.val","mean", "std.dev"),"age_at_neuro_year"]
td_age_descrip_stat <- data.frame(td_age_descrip_stat)
td_age_descrip_stat <- cbind(c("n","mean", "std.dev"), td_age_descrip_stat)
print(td_age_descrip_stat)## c("n", "mean", "std.dev") td_age_descrip_stat
## 1 n 22.000000
## 2 mean 9.284136
## 3 std.dev 2.037441asd_age_descrip_stat <- stat.desc(asd_age)
asd_age_descrip_stat <- asd_age_descrip_stat[c("nbr.val","mean", "std.dev"),"age_at_neuro_year"]
asd_age_descrip_stat <- data.frame(asd_age_descrip_stat)
asd_age_descrip_stat <- cbind(c("n","mean", "std.dev"), asd_age_descrip_stat)
print(asd_age_descrip_stat)## c("n", "mean", "std.dev") asd_age_descrip_stat
## 1 n 12.0000000
## 2 mean 7.9305833
## 3 std.dev 0.9043391#Test age difference between group, p <0.05 means age is significantly different between group
print(t.test(td_age$age_at_neuro_year, asd_age$age_at_neuro_year))##
## Welch Two Sample t-test
##
## data: td_age$age_at_neuro_year and asd_age$age_at_neuro_year
## t = 2.6708, df = 31.151, p-value = 0.01192
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## 0.3201399 2.3869662
## sample estimates:
## mean of x mean of y
## 9.284136 7.930583Age IS significantly different between TD and ASD (for all in lab participants.)
Prep data for graphs, remove outliers
iascl_sl_bar_all <- iascl_all_completed[,c(1:6, 8:11, 14:17)]
#Remove outliners for accuracy data
ssl_accurarcy_outlier <-
iascl_sl_bar_all[which(iascl_sl_bar_all$accuracy_children_ssl_accuracies < 0.15), "part_id"]
tsl_accurarcy_outlier <-
iascl_sl_bar_all[which(iascl_sl_bar_all$accuracy_children_tsl_accuracies < 0.15), "part_id"]
if(length(ssl_accurarcy_outlier) != 0) {
iascl_sl_bar_all[which(
iascl_sl_bar_all$part_id %in% ssl_accurarcy_outlier),
c("accuracy_children_ssl_accuracies")] <- NA
}
if(length(ssl_accurarcy_outlier) != 0) {
iascl_all_completed[which(
iascl_all_completed$part_id %in% ssl_accurarcy_outlier),
c("accuracy_children_ssl_accuracies")] <- NA
}
if(length(tsl_accurarcy_outlier) != 0) {
iascl_sl_bar_all[which(
iascl_sl_bar_all$part_id %in% tsl_accurarcy_outlier),
c("accuracy_children_tsl_accuracies")] <- NA
}
if(length(tsl_accurarcy_outlier) != 0) {
iascl_all_completed[which(
iascl_all_completed$part_id %in% tsl_accurarcy_outlier),
c("accuracy_children_tsl_accuracies")] <- NA
}
#Remove outliners for RT and RT slope data-----------------------------------------------------------------
ssl_rt_hit_count <- read.csv("ssl_rt_hit_trial_count.csv")
tsl_rt_hit_count <- read.csv("tsl_rt_hit_trial_count.csv")
ssl_rt_hit_count <-
ssl_rt_hit_count[which(ssl_rt_hit_count$part_id %in% iascl_sl_bar_all$part_id),]
tsl_rt_hit_count <-
tsl_rt_hit_count[which(tsl_rt_hit_count$part_id %in% iascl_sl_bar_all$part_id),]
ssl_rt_outlier <- ssl_rt_hit_count[which(ssl_rt_hit_count$hit_trial_number < 6), "part_id"]
tsl_rt_outlier <- tsl_rt_hit_count[which(tsl_rt_hit_count$hit_trial_number < 6), "part_id"]
if(length(ssl_rt_outlier) != 0) {
iascl_sl_bar_all[which(
iascl_sl_bar_all$part_id %in% ssl_rt_outlier),
c("rt_children_ssl_indiv_rts",
"slope_children_ssl_indiv_rts_slope")] <- NA
}
if(length(ssl_rt_outlier) != 0) {
iascl_all_completed[which(
iascl_all_completed$part_id %in% ssl_rt_outlier),
c("rt_children_ssl_indiv_rts",
"slope_children_ssl_indiv_rts_slope")] <- NA
}
if(length(tsl_rt_outlier) != 0) {
iascl_sl_bar_all[which(
iascl_sl_bar_all$part_id %in% tsl_rt_outlier),
c("rt_children_tsl_indiv_rts",
"slope_children_tsl_indiv_rts_slope")] <- NA
}
if(length(tsl_rt_outlier) != 0) {
iascl_all_completed[which(
iascl_all_completed$part_id %in% tsl_rt_outlier),
c("rt_children_tsl_indiv_rts",
"slope_children_tsl_indiv_rts_slope")] <- NA
}These are removed outliers:
SSL RT and Slope < 6 hits:
TSL RT and Slope < 6 hits: blast_c_209
SSL Accuracy:
TSL Accuracy:
Group SL analysis & KBITT, SCQ Descriptive Stat
T-test TD vs. ASD
td_sl <-
iascl_all_completed[which(iascl_all_completed$group == "TD"),]
asd_sl <-
iascl_all_completed[which(iascl_all_completed$group == "ASD"),]
library("pastecs")
td_kbit_scq_descrip_stat <- stat.desc(td_sl[,c(3:4,5)])
td_kbit_scq_descrip_stat <- td_kbit_scq_descrip_stat[c("nbr.val","mean", "std.dev"),
c("kbit_matrices_raw", "kbit_matrices_std", "scq_total")]
print(td_kbit_scq_descrip_stat)## kbit_matrices_raw kbit_matrices_std scq_total
## nbr.val 22.000000 22.00000 20.000000
## mean 31.409091 113.18182 2.150000
## std.dev 6.870635 16.01379 1.565248asd_kbit_scq_descrip_stat <- stat.desc(asd_sl[,c(3:4,5)])
asd_kbit_scq_descrip_stat <- asd_kbit_scq_descrip_stat[c("nbr.val","mean", "std.dev"),
c("kbit_matrices_raw", "kbit_matrices_std", "scq_total")]
print(asd_kbit_scq_descrip_stat)## kbit_matrices_raw kbit_matrices_std scq_total
## nbr.val 12.000000 12.00000 11.000000
## mean 26.333333 106.91667 13.909091
## std.dev 8.370221 18.54948 5.048852#p <0.05, TD and ASD are truly different
#t.test(td_sl$accuracy_children_ssl_accuracies, asd_sl$accuracy_children_ssl_accuracies)
t_test_multi_pair <-
function(x,y){
test <- t.test(x,y)
data.frame(p_value = test$p.value,
df = test$parameter,
t_stat = test$statistic)
}
#Results withOUT outliers
sapply(intersect(colnames(td_sl),colnames(asd_sl))[c(8:11, 14:17)],
function(x) t_test_multi_pair(td_sl[,x], asd_sl[,x]))## accuracy_children_ssl_accuracies accuracy_children_tsl_accuracies
## p_value 0.5655127 0.135311
## df 17.27191 24.96517
## t_stat 0.5859128 1.543421
## entropy_children_ssl_entropy entropy_children_tsl_entropy
## p_value 0.9006048 0.24506
## df 20.72792 18.14556
## t_stat -0.1264365 -1.201355
## rt_children_ssl_indiv_rts rt_children_tsl_indiv_rts
## p_value 0.02945021 0.2593682
## df 15.33708 23.84279
## t_stat 2.400955 1.155416
## slope_children_ssl_indiv_rts_slope
## p_value 0.1201148
## df 16.13298
## t_stat -1.641148
## slope_children_tsl_indiv_rts_slope
## p_value 0.8822999
## df 19.88359
## t_stat -0.149969sapply(intersect(colnames(td_sl),colnames(asd_sl))[c(3:4, 5)],
function(x) t_test_multi_pair(td_sl[,x], asd_sl[,x]))## kbit_matrices_raw kbit_matrices_std scq_total
## p_value 0.08818368 0.3356349 1.117523e-05
## df 19.2119 20.02764 11.0689
## t_stat 1.796338 0.9865457 -7.52821No group difference is seen for KBIT.
SSL Mean RT is significantly different between group (TD > ASD). No outliners in the in lab sample.
SCQ significantly higher in ASD.
Ancova comparing between group for all measures
ancova_bt_group <-
function(x,y,z){
test <- aov(x~y + age_at_neuro_year, z)
summary(test)
}
#Results with outliers
sapply(colnames(iascl_all_completed)[c(8:11, 14:17)],
function(x) ancova_bt_group(iascl_all_completed[,x], iascl_all_completed[,"group"], iascl_all_completed))## $accuracy_children_ssl_accuracies
## Df Sum Sq Mean Sq F value Pr(>F)
## y 1 0.003110 0.0031097 0.3363 0.5684
## age_at_neuro_year 1 0.005692 0.0056921 0.6156 0.4419
## Residuals 20 0.184924 0.0092462
##
## $accuracy_children_tsl_accuracies
## Df Sum Sq Mean Sq F value Pr(>F)
## y 1 0.03767 0.037673 2.2743 0.14459
## age_at_neuro_year 1 0.05279 0.052792 3.1870 0.08687 .
## Residuals 24 0.39755 0.016565
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## $entropy_children_ssl_entropy
## Df Sum Sq Mean Sq F value Pr(>F)
## y 1 0.000163 0.0001633 0.0131 0.9101
## age_at_neuro_year 1 0.000030 0.0000299 0.0024 0.9614
## Residuals 20 0.249840 0.0124920
##
## $entropy_children_tsl_entropy
## Df Sum Sq Mean Sq F value Pr(>F)
## y 1 0.03307 0.033065 1.2618 0.27242
## age_at_neuro_year 1 0.16940 0.169400 6.4644 0.01787 *
## Residuals 24 0.62892 0.026205
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## $rt_children_ssl_indiv_rts
## Df Sum Sq Mean Sq F value Pr(>F)
## y 1 162142 162142 5.9346 0.02432 *
## age_at_neuro_year 1 6211 6211 0.2273 0.63867
## Residuals 20 546428 27321
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## $rt_children_tsl_indiv_rts
## Df Sum Sq Mean Sq F value Pr(>F)
## y 1 37580 37580 1.2691 0.2715
## age_at_neuro_year 1 47880 47880 1.6170 0.2162
## Residuals 23 681049 29611
##
## $slope_children_ssl_indiv_rts_slope
## Df Sum Sq Mean Sq F value Pr(>F)
## y 1 0.0015102 0.00151017 2.8766 0.1054
## age_at_neuro_year 1 0.0008359 0.00083595 1.5923 0.2215
## Residuals 20 0.0104998 0.00052499
##
## $slope_children_tsl_indiv_rts_slope
## Df Sum Sq Mean Sq F value Pr(>F)
## y 1 0.0000103 0.00001028 0.0225 0.8820
## age_at_neuro_year 1 0.0000105 0.00001052 0.0230 0.8807
## Residuals 23 0.0104977 0.00045642Ancova controlling for age shows two groups are still significantly different in SSL mean RT. Also, Age had an significant effect on TSL entropy (F = 6.4644, p = 0.01787). Age on TSL accuracy (F = 3.1870, p = 0.08687)
Accuracy against chance for TD and ASD
t_test_against_chance <-
function(x){
test <- t.test(x, mu=0.5, alternative= "greater")
data.frame(p_value = test$p.value,
df = test$parameter,
t_stat = test$statistic)
}
#TD without outliers
sapply(colnames(td_sl)[c(8:9)],
function(x) t_test_against_chance(td_sl[,x]))## accuracy_children_ssl_accuracies accuracy_children_tsl_accuracies
## p_value 0.0515594 0.0111262
## df 13 15
## t_stat 1.753091 2.548765#ASD without outliers
sapply(colnames(asd_sl)[c(8:9)],
function(x) t_test_against_chance(asd_sl[,x]))## accuracy_children_ssl_accuracies accuracy_children_tsl_accuracies
## p_value 0.2640249 0.2747999
## df 8 10
## t_stat 0.6595811 0.6192483TD: marginally significant above chance for SSL accuracy (t = 1.75, p = 0.051) Significantly above chance for TSL accuracy (t = 2.55, p = 0.011)
ASD: Not above chance for either.
Ignore: Accuracy bar graph without outliers
Accuracy Line Plot (no outliers)
detach(package:plyr)
library(dplyr)
library(ggplot2)
#The group_by is strange in dplyr, n keeps changing when I change mutate labels
# iascl_sl_bar_acc %>%
# filter(value != "") %>%
# mutate(group = factor(group, labels = c("ASD", "TD")),
# variable = factor(variable, labels = c("Tone", "Syllable"))) %>%
# group_by(group, variable) %>%
# summarise(acc_mean = mean(value, na.rm = TRUE),
# acc_se = sd(value, na.rm = TRUE)/ sqrt(n()),
# n = n())
#Descriptive Stats for TD Tone Accuracy
library("pastecs")
stat_desc <-
function(df,group,task) {
stat.desc(df[which(df$group == group & df$variable == task),])
}
td_tone <- stat_desc(iascl_sl_bar_acc, "TD", "Tone")
td_syllable <- stat_desc(iascl_sl_bar_acc, "TD", "Syllable")
asd_tone <- stat_desc(iascl_sl_bar_acc, "ASD", "Tone")
asd_syllable <- stat_desc(iascl_sl_bar_acc, "ASD", "Syllable")
extract_line_graph_stat <- function (x, group, task) {
y <- x[c("nbr.val","mean", "SE.mean"),"value"]
dim(y) <- c(3, 1)
y <- t(y)
colnames(y) <- c("n", "mean", "std_error")
y <- data.frame(y)
y$group <- group
y$task <- task
return(y)
}
td_tone_stat <- extract_line_graph_stat(td_tone, "TD", "Tone")
td_syllable_stat <- extract_line_graph_stat(td_syllable, "TD", "Syllable")
asd_tone_stat <- extract_line_graph_stat(asd_tone, "ASD", "Tone")
asd_syllable_stat <- extract_line_graph_stat(asd_syllable, "ASD", "Syllable")
acc_stat <- rbind(td_tone_stat, td_syllable_stat, asd_tone_stat, asd_syllable_stat)
colnames(acc_stat)[colnames(acc_stat) == "group"] <- "Group"
library(plyr)## -------------------------------------------------------------------------## You have loaded plyr after dplyr - this is likely to cause problems.
## If you need functions from both plyr and dplyr, please load plyr first, then dplyr:
## library(plyr); library(dplyr)## -------------------------------------------------------------------------##
## Attaching package: 'plyr'## The following objects are masked from 'package:dplyr':
##
## arrange, count, desc, failwith, id, mutate, rename, summarise,
## summarize## The following objects are masked from 'package:reshape':
##
## rename, round_anyacc_stat$task <-
revalue(acc_stat$task, c(#"entropy_children_lsl_entropy"="Letter",
#"entropy_children_vsl_entropy"="Image",
"Syllable"="Linguistic",
"Tone"="Non-linguistic"))
pd <- position_dodge(width = 0.2)
acc_stat %>%
ggplot(aes(x = task, y = mean, group = Group)) +
geom_line(aes(linetype = Group, color = Group), position = pd, size = 1.8) +
geom_errorbar(aes(ymin = mean - std_error, ymax = mean + std_error), width = .1, position = pd) +
geom_point(aes(color = Group), size = 4, position = pd,show.legend = FALSE) +
geom_point(size = 3, color = "white", position = pd) +
labs(title = "Behavioral Accuracy across Groups",
x = "SL Task", # Change x-axis label
y = "Mean Accuracy") +
theme(plot.title = element_text(hjust = 0.5)) +
theme(
plot.title = element_text(size=16, face="bold"),
axis.title.x = element_text(size=14, face="bold"),
axis.title.y = element_text(size=14, face="bold"),
axis.text=element_text(size=12, face = "bold")
) +
theme(legend.text=element_text(size=14, face="bold"),
legend.title=element_text(size=15, face="bold")) +
theme(
panel.background = element_rect(fill = "white"), # Set plot background to white
legend.key = element_rect(fill = "white"), # Set legend item backgrounds to white
axis.line.x = element_line(colour = "black", size = 1), # Add line to x axis
axis.line.y = element_line(colour = "black", size = 1) # Add line to y axis
) +
geom_hline(yintercept=0.5, color = "grey") +
geom_text(aes(x= 2.3, y = 0.505, label = paste0("chance", "-level")), size=5) +
geom_text(aes(x= 2.05, y = 0.64, label = paste0("*")), size=7) +
geom_text(aes(x= 1.05, y = 0.64, label = paste0("†")), size=7) 
# ggsave("behavioral_acc_across_group.png",
# width = 15, height = 15, units = "cm")Ignore: Entropy bar graph without outliers
Entropy Line Plot (no outliers)
detach(package:plyr)
library(dplyr)
library(ggplot2)
library("pastecs")
stat_desc <-
function(df,group,task) {
stat.desc(df[which(df$group == group & df$variable == task),])
}
td_tone <- stat_desc(iascl_sl_bar_entropy, "TD", "Tone")
td_syllable <- stat_desc(iascl_sl_bar_entropy, "TD", "Syllable")
asd_tone <- stat_desc(iascl_sl_bar_entropy, "ASD", "Tone")
asd_syllable <- stat_desc(iascl_sl_bar_entropy, "ASD", "Syllable")
extract_line_graph_stat <- function (x, group, task) {
y <- x[c("nbr.val","mean", "SE.mean"),"value"]
dim(y) <- c(3, 1)
y <- t(y)
colnames(y) <- c("n", "mean", "std_error")
y <- data.frame(y)
y$group <- group
y$task <- task
return(y)
}
td_tone_stat <- extract_line_graph_stat(td_tone, "TD", "Tone")
td_syllable_stat <- extract_line_graph_stat(td_syllable, "TD", "Syllable")
asd_tone_stat <- extract_line_graph_stat(asd_tone, "ASD", "Tone")
asd_syllable_stat <- extract_line_graph_stat(asd_syllable, "ASD", "Syllable")
entropy_stat <- rbind(td_tone_stat, td_syllable_stat, asd_tone_stat, asd_syllable_stat)
pd <- position_dodge(width = 0.2)
entropy_stat %>%
ggplot(aes(x = task, y = mean, group = group)) +
geom_line(aes(linetype = group), position = pd) +
geom_errorbar(aes(ymin = mean - std_error, ymax = mean + std_error), width = .1, position = pd) +
geom_point(size = 4, position = pd) +
geom_point(size = 3, color = "white", position = pd) +
guides(linetype = guide_legend("Group")) +
labs(title = "Behavioral SL Task Entropy",
x = "Task", # Change x-axis label
y = "Entropy") +
theme(plot.title = element_text(hjust = 0.5)) + #Center title
theme(
panel.background = element_rect(fill = "white"), # Set plot background to white
legend.key = element_rect(fill = "white"), # Set legend item backgrounds to white
axis.line.x = element_line(colour = "black", size = 1), # Add line to x axis
axis.line.y = element_line(colour = "black", size = 1) # Add line to y axis
) 
Mean RT bar graph without outliers
Mean RT Line Plot (no outliers)
detach(package:plyr)
library(dplyr)
library(ggplot2)
library("pastecs")
stat_desc <-
function(df,group,task) {
stat.desc(df[which(df$group == group & df$variable == task),])
}
td_tone <- stat_desc(iascl_sl_bar_rt, "TD", "Tone")
td_syllable <- stat_desc(iascl_sl_bar_rt, "TD", "Syllable")
asd_tone <- stat_desc(iascl_sl_bar_rt, "ASD", "Tone")
asd_syllable <- stat_desc(iascl_sl_bar_rt, "ASD", "Syllable")
extract_line_graph_stat <- function (x, group, task) {
y <- x[c("nbr.val","mean", "SE.mean"),"value"]
dim(y) <- c(3, 1)
y <- t(y)
colnames(y) <- c("n", "mean", "std_error")
y <- data.frame(y)
y$group <- group
y$task <- task
return(y)
}
td_tone_stat <- extract_line_graph_stat(td_tone, "TD", "Tone")
td_syllable_stat <- extract_line_graph_stat(td_syllable, "TD", "Syllable")
asd_tone_stat <- extract_line_graph_stat(asd_tone, "ASD", "Tone")
asd_syllable_stat <- extract_line_graph_stat(asd_syllable, "ASD", "Syllable")
rt_stat <- rbind(td_tone_stat, td_syllable_stat, asd_tone_stat, asd_syllable_stat)
pd <- position_dodge(width = 0.2)
rt_stat %>%
ggplot(aes(x = task, y = mean, group = group)) +
geom_line(aes(linetype = group), position = pd) +
geom_errorbar(aes(ymin = mean - std_error, ymax = mean + std_error), width = .1, position = pd) +
geom_point(size = 4, position = pd) +
geom_point(size = 3, color = "white", position = pd) +
guides(linetype = guide_legend("Group")) +
labs(title = "Behavioral SL Task Mean Reaction Time",
x = "Task", # Change x-axis label
y = "Mean Reaction Time (ms)") +
theme(plot.title = element_text(hjust = 0.5)) + #Center title
theme(
panel.background = element_rect(fill = "white"), # Set plot background to white
legend.key = element_rect(fill = "white"), # Set legend item backgrounds to white
axis.line.x = element_line(colour = "black", size = 1), # Add line to x axis
axis.line.y = element_line(colour = "black", size = 1) # Add line to y axis
) 
Scaled RT Slope bar graph without outliers
RT Slope Line Plot (no outliers)
detach(package:plyr)
library(dplyr)
library(ggplot2)
library("pastecs")
stat_desc <-
function(df,group,task) {
stat.desc(df[which(df$group == group & df$variable == task),])
}
td_tone <- stat_desc(iascl_sl_bar_slope, "TD", "Tone")
td_syllable <- stat_desc(iascl_sl_bar_slope, "TD", "Syllable")
asd_tone <- stat_desc(iascl_sl_bar_slope, "ASD", "Tone")
asd_syllable <- stat_desc(iascl_sl_bar_slope, "ASD", "Syllable")
extract_line_graph_stat <- function (x, group, task) {
y <- x[c("nbr.val","mean", "SE.mean"),"value"]
dim(y) <- c(3, 1)
y <- t(y)
colnames(y) <- c("n", "mean", "std_error")
y <- data.frame(y)
y$group <- group
y$task <- task
return(y)
}
td_tone_stat <- extract_line_graph_stat(td_tone, "TD", "Tone")
td_syllable_stat <- extract_line_graph_stat(td_syllable, "TD", "Syllable")
asd_tone_stat <- extract_line_graph_stat(asd_tone, "ASD", "Tone")
asd_syllable_stat <- extract_line_graph_stat(asd_syllable, "ASD", "Syllable")
slope_stat <- rbind(td_tone_stat, td_syllable_stat, asd_tone_stat, asd_syllable_stat)
pd <- position_dodge(width = 0.2)
colnames(slope_stat)[colnames(slope_stat) == "group"] <- "Group"
library(plyr)## -------------------------------------------------------------------------## You have loaded plyr after dplyr - this is likely to cause problems.
## If you need functions from both plyr and dplyr, please load plyr first, then dplyr:
## library(plyr); library(dplyr)## -------------------------------------------------------------------------##
## Attaching package: 'plyr'## The following objects are masked from 'package:dplyr':
##
## arrange, count, desc, failwith, id, mutate, rename, summarise,
## summarize## The following objects are masked from 'package:reshape':
##
## rename, round_anyslope_stat$task <-
revalue(slope_stat$task, c(#"entropy_children_lsl_entropy"="Letter",
#"entropy_children_vsl_entropy"="Image",
"Syllable"="Linguistic",
"Tone"="Non-linguistic"))
pd <- position_dodge(width = 0.2)
slope_stat %>%
ggplot(aes(x = task, y = mean, group = Group)) +
geom_line(aes(linetype = Group, color = Group), position = pd, size = 1.8) +
geom_errorbar(aes(ymin = mean - std_error, ymax = mean + std_error), width = .1, position = pd) +
geom_point(aes(color = Group), size = 4, position = pd,show.legend = FALSE) +
geom_point(size = 3, color = "white", position = pd) +
labs(title = "Behavioral Reaction Time Slope across Groups",
x = "SL Task", # Change x-axis label
y = "Reaction Time Slope (arbitrary unit/trial)") +
theme(plot.title = element_text(hjust = 0.5)) +
theme(
plot.title = element_text(size=16, face="bold"),
axis.title.x = element_text(size=14, face="bold"),
axis.title.y = element_text(size=14, face="bold"),
axis.text=element_text(size=12, face = "bold")
) +
theme(legend.text=element_text(size=14, face="bold"),
legend.title=element_text(size=15, face="bold")) +
theme(
panel.background = element_rect(fill = "white"), # Set plot background to white
legend.key = element_rect(fill = "white"), # Set legend item backgrounds to white
axis.line.x = element_line(colour = "black", size = 1), # Add line to x axis
axis.line.y = element_line(colour = "black", size = 1) # Add line to y axis
) +
geom_text(aes(x= 1, y = 0.015, label = paste0("†")), size=7) 
# ggsave(
# "behavioral_slope_across_group.png",
# width = 15, height = 15, units = "cm"
# )Behavioral Correlations
Heat Map
Heat Map for R, P, N for behavioral correlations
Within Task, Across Measures
#GGplot the heat map----------------------------------------------------------------------------------------------
ggplot(data = heat_map_corr_data, aes(x = column, y = row, fill = cor)) +
geom_tile(color = "white") +
scale_fill_gradient2(
low = "blue",
high = "red",
mid = "white",
midpoint = 0,
limit = c(-1, 1),
space = "Lab",
name = "pearson\nCorrelation r"
) +
theme_minimal() +
theme(axis.text.x = element_text(
angle = 90,
vjust = 1,
size = 12,
hjust = 1
)) +
theme(axis.title.x = element_blank(),
axis.text.y = element_text (size = 12),
axis.title.y = element_blank()) +
coord_fixed() +
geom_text(aes(label = cor), color = "black", size = 2)
ggplot(data = heat_map_corr_data, aes(x = column, y = row, fill = p)) +
geom_tile(color = "white") +
scale_fill_gradient2(
low = "blue",
high = "red",
mid = "white",
midpoint = 0,
limit = c(0, 1),
space = "Lab",
name = "p-value"
) +
theme_minimal() +
theme(axis.text.x = element_text(
angle = 90,
vjust = 1,
size = 12,
hjust = 1
)) +
theme(axis.title.x = element_blank(),
axis.text.y = element_text (size = 12),
axis.title.y = element_blank()) +
coord_fixed() +
geom_text(aes(label = p), color = "black", size = 2)
ggplot(data = heat_map_corr_data, aes(x = column, y = row, fill = n)) +
geom_tile(color = "white") +
scale_fill_gradient2(
low = "blue",
high = "red",
mid = "white",
midpoint = 0,
limit = c(0, 33),
space = "Lab",
name = "n"
) +
theme_minimal() +
theme(axis.text.x = element_text(
angle = 90,
vjust = 1,
size = 12,
hjust = 1
)) +
theme(axis.title.x = element_blank(),
axis.text.y = element_text (size = 12),
axis.title.y = element_blank()) +
coord_fixed() +
geom_text(aes(label = n), color = "black", size = 3.5)
Across Behavioral Tasks, Within Measure
ggplot(data = heat_map_corr_data_task, aes(x = column, y = row, fill = cor)) +
geom_tile(color = "white") +
scale_fill_gradient2(
low = "blue",
high = "red",
mid = "white",
midpoint = 0,
limit = c(-1, 1),
space = "Lab",
name = "pearson\nCorrelation r"
) +
theme_minimal() +
theme(axis.text.x = element_text(
angle = 90,
vjust = 1,
size = 12,
hjust = 1
)) +
theme(axis.title.x = element_blank(),
axis.text.y = element_text (size = 12),
axis.title.y = element_blank()) +
coord_fixed() +
geom_text(aes(label = cor), color = "black", size = 2)
ggplot(data = heat_map_corr_data_task, aes(x = column, y = row, fill = p)) +
geom_tile(color = "white") +
scale_fill_gradient2(
low = "blue",
high = "red",
mid = "white",
midpoint = 0,
limit = c(0, 1),
space = "Lab",
name = "p-value"
) +
theme_minimal() +
theme(axis.text.x = element_text(
angle = 90,
vjust = 1,
size = 12,
hjust = 1
)) +
theme(axis.title.x = element_blank(),
axis.text.y = element_text (size = 12),
axis.title.y = element_blank()) +
coord_fixed() +
geom_text(aes(label = p), color = "black", size = 2)
Beh Corr Plots
ggplot(data = iascl_all_completed,
aes(y= gjt_std_score_gjt_web,
x = scq_total,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2") +
guides(linetype = guide_legend("Group")) +
labs(title = "Grammaticality Judgement Task and Social Communication Questionnaire",
y = "Grammaticality Judgement Standardized Score", # Change x-axis label
x = "Social Communication Questionnaire") +
theme(plot.title = element_text(hjust = 0.5)) +
geom_text(aes(x = 20, y = 120, label = paste("rho =", "-0.76**")),color = "red",size=4)## Warning: Removed 21 rows containing missing values (geom_point).
ggplot(data = iascl_all_completed,
aes(y= a_prime,
x = scq_total,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2") +
guides(linetype = guide_legend("Group")) +
labs(title = "Grammaticality Judgement Task and Social Communication Questionnaire",
y = "Grammaticality Judgement Raw Score", # Change x-axis label
x = "Social Communication Questionnaire") +
theme(plot.title = element_text(hjust = 0.5)) +
geom_text(aes(x = 20, y = 1, label = paste("rho =", "-0.65*")),color = "red",size=4)## Warning: Removed 21 rows containing missing values (geom_point).
ggplot(data = iascl_all_completed,
aes(x= a_prime,
y = scq_total,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2") +
guides(linetype = guide_legend("Group")) +
labs(title = "Grammaticality Judgement Task and Social Communication Questionnaire",
x = "Grammaticality Judgement Raw Score", # Change x-axis label
y = "Social Communication Questionnaire") +
theme(plot.title = element_text(hjust = 0.5)) +
geom_text(aes(y = 20, x = 0.95, label = paste("rho =", "-0.65*")),color = "red",size=4)## Warning: Removed 21 rows containing missing values (geom_point).
ggplot(data = iascl_all_completed,
aes(x= a_prime,
y = scq_total,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2") +
guides(linetype = guide_legend("Group")) +
labs(title = "Grammaticality Judgement Task and Social Communication Questionnaire",
x = "Grammaticality Judgement Raw Score", # Change x-axis label
y = "Social Communication Questionnaire") +
theme(plot.title = element_text(hjust = 0.5)) +
geom_text(aes(y = 20, x = 0.95, label = paste("rho =", "-0.65*")),color = "red",size=4)## Warning: Removed 21 rows containing missing values (geom_point).
Ignore: ASD Beh Corr Plot
#MRI T-test between group
iascl_beh_gray_matter_td[,c(2:10, 13:23)] <-
lapply(iascl_beh_gray_matter_td[,c(2:10, 13:23)], function(x) {
if (is.factor(x))
as.numeric(as.character(x))
else
x
})
iascl_beh_gray_matter_asd[,c(2:10, 13:23)] <-
lapply(iascl_beh_gray_matter_asd[,c(2:10, 13:23)], function(x) {
if (is.factor(x))
as.numeric(as.character(x))
else
x
})
iascl_beh_cor_thick_td[,c(2:10, 13:23)] <-
lapply(iascl_beh_cor_thick_td[,c(2:10, 13:23)], function(x) {
if (is.factor(x))
as.numeric(as.character(x))
else
x
})
iascl_beh_cor_thick_asd[,c(2:10, 13:23)] <-
lapply(iascl_beh_cor_thick_asd[,c(2:10, 13:23)], function(x) {
if (is.factor(x))
as.numeric(as.character(x))
else
x
})
iascl_beh_gray_matter[,c(2:10, 13:23)] <-
lapply(iascl_beh_gray_matter[,c(2:10, 13:23)], function(x) {
if (is.factor(x))
as.numeric(as.character(x))
else
x
})
iascl_beh_cor_thick[,c(2:10, 13:23)] <-
lapply(iascl_beh_cor_thick[,c(2:10, 13:23)], function(x) {
if (is.factor(x))
as.numeric(as.character(x))
else
x
})
#T test between group
t_test_multi_pair <-
function(x,y){
test <- t.test(x,y)
data.frame(p_value = test$p.value,
df = test$parameter,
t_stat = test$statistic)
}
# GM T-test between group
sapply(intersect(colnames(iascl_beh_gray_matter_td),colnames(iascl_beh_gray_matter_asd))[2:10],
function(x) t_test_multi_pair(iascl_beh_gray_matter_td[,x], iascl_beh_gray_matter_asd[,x]))## LeftAngG LeftIFG LeftIFGorb LeftMFG LeftMidAntTemp
## p_value 0.3709168 0.2490072 0.3570282 0.2734187 0.4028914
## df 5.571482 5.095522 4.784659 3.941682 3.900906
## t_stat -0.9729214 -1.300917 -1.018772 -1.271432 -0.9373715
## LeftMidPostTemp LeftPostTemp RightMidAntTemp RightMidPostTemp
## p_value 0.2847049 0.6574654 0.3053455 0.1005669
## df 6.152649 8.325675 3.941514 9.83297
## t_stat -1.171644 -0.4597477 -1.177003 -1.81211# CT T-test between group
sapply(intersect(colnames(iascl_beh_cor_thick_td),colnames(iascl_beh_cor_thick_asd))[2:10],
function(x) t_test_multi_pair(iascl_beh_cor_thick_td[,x], iascl_beh_cor_thick_asd[,x]))## LeftAngG LeftIFG LeftIFGorb LeftMFG LeftMidAntTemp
## p_value 0.1295935 0.0905683 0.4910804 0.07046103 0.9654963
## df 8.541535 10.5116 3.995009 4.960533 12.9725
## t_stat -1.677366 -1.863315 -0.7573 -2.29689 0.0441
## LeftMidPostTemp LeftPostTemp RightMidAntTemp RightMidPostTemp
## p_value 0.1718029 0.1839262 0.6907809 0.2909642
## df 5.926682 7.287599 5.010944 7.925119
## t_stat -1.553923 -1.467877 -0.4216162 -1.131385# ETIV T-test between groupNo significant difference between TD and ASD for Cortical thickness or Gray Matter.
ETIV-Gray Matter Corr (Only 13 have ETIV + GM)
etiv <-
read.csv("/Users/jojohu/Documents/Qlab/iascl/aseg_stats_etiv_measure.csv")
library("stringr")
etiv$Measure.volume <- str_extract(etiv$Measure.volume, "(?<=sub-)\\S{9}(?=/)")
etiv$Measure.volume ## [1] "blasta001" "blasta002" "blasta004" "blasta005" "blasta006"
## [6] "blasta007" "blasta008" "blasta010" "blasta011" "blasta012"
## [11] "blasta013" "blasta015" "blasta017" "blasta018" "blasta020"
## [16] "blasta021" "blasta022" "blasta023" "blasta024" "blasta025"
## [21] "blasta026" "blasta027" "blasta028" "blasta029" "blastc009"
## [26] "blastc019" "blastc059" "blastc061" "blastc071" "blastc095"
## [31] "blastc126" "blastc147" "blastc165" "blastc168" "blastc206"
## [36] "blastc210" "blastc213" "blastc224"partid1 <- substr(etiv$Measure.volume, start = 1, stop = 5)
partid2 <- substr(etiv$Measure.volume, start = 6, stop = 6)
partid3 <- substr(etiv$Measure.volume, start = 7, stop = 9)
etiv$part_id <- paste(partid1, partid2, partid3, sep = "_")
mri_etiv_gm_wide <-
merge(mri_gray_matter_wide,
etiv[,c("part_id", "EstimatedTotalIntraCranialVol")],
by.x = "part_id", by.y = "part_id",
all.x = TRUE)
etiv_with_group <-
merge(iascl_beh_cor_thick, etiv[,c("part_id", "EstimatedTotalIntraCranialVol")],
by.x = "part_id", by.y="part_id",
all.x = TRUE)
#T test for ETIV
t.test(etiv_with_group[which(etiv_with_group$group == "TD"), "EstimatedTotalIntraCranialVol"],
etiv_with_group[which(etiv_with_group$group == "ASD"), "EstimatedTotalIntraCranialVol"])##
## Welch Two Sample t-test
##
## data: etiv_with_group[which(etiv_with_group$group == "TD"), "EstimatedTotalIntraCranialVol"] and etiv_with_group[which(etiv_with_group$group == "ASD"), "EstimatedTotalIntraCranialVol"]
## t = -1.5014, df = 10.338, p-value = 0.1632
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -248036.55 47803.49
## sample estimates:
## mean of x mean of y
## 1549337 1649453mri_etiv_gm_corr <- rcorr(as.matrix(
mri_etiv_gm_wide[,c(2:11)]), type = c("pearson"))
mri_etiv_gm_pearson_corr <-
flattenCorrMatrix(mri_etiv_gm_corr$r, mri_etiv_gm_corr$P, mri_etiv_gm_corr$n)
mri_etiv_gm_pearson_corr[,c(3:5)] <-
round(mri_etiv_gm_pearson_corr[,c(3:5)], 3)
mri_etiv_gm_pearson_corr <-
mri_etiv_gm_pearson_corr[which(mri_etiv_gm_pearson_corr$p <= 0.07),]
print(mri_etiv_gm_pearson_corr)## row column cor p n
## 1 LeftAngG LeftIFG 0.813 0.000 16
## 2 LeftAngG LeftIFGorb 0.849 0.000 16
## 3 LeftIFG LeftIFGorb 0.847 0.000 16
## 4 LeftAngG LeftMFG 0.688 0.003 16
## 5 LeftIFG LeftMFG 0.858 0.000 16
## 6 LeftIFGorb LeftMFG 0.695 0.003 16
## 8 LeftIFG LeftMidAntTemp 0.464 0.070 16
## 9 LeftIFGorb LeftMidAntTemp 0.559 0.024 16
## 10 LeftMFG LeftMidAntTemp 0.714 0.002 16
## 11 LeftAngG LeftMidPostTemp 0.706 0.002 16
## 12 LeftIFG LeftMidPostTemp 0.733 0.001 16
## 13 LeftIFGorb LeftMidPostTemp 0.639 0.008 16
## 14 LeftMFG LeftMidPostTemp 0.577 0.019 16
## 16 LeftAngG LeftPostTemp 0.643 0.007 16
## 17 LeftIFG LeftPostTemp 0.833 0.000 16
## 18 LeftIFGorb LeftPostTemp 0.673 0.004 16
## 19 LeftMFG LeftPostTemp 0.706 0.002 16
## 21 LeftMidPostTemp LeftPostTemp 0.655 0.006 16
## 25 LeftMFG RightMidAntTemp 0.480 0.060 16
## 26 LeftMidAntTemp RightMidAntTemp 0.619 0.011 16
## 28 LeftPostTemp RightMidAntTemp 0.530 0.035 16
## 29 LeftAngG RightMidPostTemp 0.515 0.041 16
## 30 LeftIFG RightMidPostTemp 0.631 0.009 16
## 32 LeftMFG RightMidPostTemp 0.655 0.006 16
## 34 LeftMidPostTemp RightMidPostTemp 0.483 0.058 16
## 35 LeftPostTemp RightMidPostTemp 0.613 0.012 16
## 36 RightMidAntTemp RightMidPostTemp 0.496 0.051 16
## 37 LeftAngG EstimatedTotalIntraCranialVol 0.696 0.008 13
## 38 LeftIFG EstimatedTotalIntraCranialVol 0.694 0.009 13
## 39 LeftIFGorb EstimatedTotalIntraCranialVol 0.773 0.002 13
## 40 LeftMFG EstimatedTotalIntraCranialVol 0.568 0.043 13
## 41 LeftMidAntTemp EstimatedTotalIntraCranialVol 0.555 0.049 13
## 42 LeftMidPostTemp EstimatedTotalIntraCranialVol 0.581 0.037 13
## 45 RightMidPostTemp EstimatedTotalIntraCranialVol 0.700 0.008 13ETIV not significantly different between group.
LeftAngG, LeftIFG, LeftIFGorb, LeftMFG, LeftMidAntTemp are correlated with ETIV.
ETIV-Cortical Thickness Correlation (only 13 have ETIV plus Cortical Thickness)
mri_etiv_ct_wide <-
merge(average_cor_thick_wide,
etiv[,c("part_id", "EstimatedTotalIntraCranialVol")],
by.x = "part_id", by.y = "part_id",
all.x = TRUE)
mri_etiv_ct_corr <- rcorr(as.matrix(
mri_etiv_ct_wide[,c(2:11)]), type = c("pearson"))
mri_etiv_ct_pearson_corr <-
flattenCorrMatrix(mri_etiv_ct_corr$r, mri_etiv_ct_corr$P, mri_etiv_ct_corr$n)
mri_etiv_ct_pearson_corr[,c(3:5)] <-
round(mri_etiv_ct_pearson_corr[,c(3:5)], 3)
mri_etiv_ct_pearson_corr <-
mri_etiv_ct_pearson_corr[which(mri_etiv_ct_pearson_corr$p <= 0.07),]
print(mri_etiv_ct_pearson_corr)## row column cor p n
## 1 LeftAngG LeftIFG 0.499 0.049 16
## 3 LeftIFG LeftIFGorb 0.659 0.006 16
## 4 LeftAngG LeftMFG 0.635 0.008 16
## 5 LeftIFG LeftMFG 0.772 0.000 16
## 12 LeftIFG LeftMidPostTemp 0.754 0.001 16
## 13 LeftIFGorb LeftMidPostTemp 0.469 0.067 16
## 14 LeftMFG LeftMidPostTemp 0.683 0.004 16
## 15 LeftMidAntTemp LeftMidPostTemp 0.581 0.018 16
## 16 LeftAngG LeftPostTemp 0.552 0.027 16
## 17 LeftIFG LeftPostTemp 0.587 0.017 16
## 19 LeftMFG LeftPostTemp 0.733 0.001 16
## 21 LeftMidPostTemp LeftPostTemp 0.488 0.055 16
## 29 LeftAngG RightMidPostTemp 0.518 0.040 16
## 30 LeftIFG RightMidPostTemp 0.522 0.038 16
## 31 LeftIFGorb RightMidPostTemp 0.532 0.034 16Cortical thickness is not correlated with ETIV.
MRI-Behavioral Corr (ASD + TD without behavioral outliers)
Gray Matter-Beh Corr (ASD + TD)
iascl_beh_gray_matter_all <- rcorr(as.matrix(
iascl_beh_gray_matter[,c(2:10, 13:23)]), type = c("pearson"))
iascl_beh_gray_matter_pearson_all <-
flattenCorrMatrix(iascl_beh_gray_matter_all$r, iascl_beh_gray_matter_all$P, iascl_beh_gray_matter_all$n)
iascl_beh_gray_matter_pearson_all[,c(3:5)] <-
round(iascl_beh_gray_matter_pearson_all[,c(3:5)], 3)
iascl_beh_gray_matter_pearson_all <-
iascl_beh_gray_matter_pearson_all[which(iascl_beh_gray_matter_pearson_all$p <= 0.05),]
print(iascl_beh_gray_matter_pearson_all)## row column
## 1 LeftAngG LeftIFG
## 2 LeftAngG LeftIFGorb
## 3 LeftIFG LeftIFGorb
## 4 LeftAngG LeftMFG
## 5 LeftIFG LeftMFG
## 6 LeftIFGorb LeftMFG
## 9 LeftIFGorb LeftMidAntTemp
## 10 LeftMFG LeftMidAntTemp
## 11 LeftAngG LeftMidPostTemp
## 12 LeftIFG LeftMidPostTemp
## 13 LeftIFGorb LeftMidPostTemp
## 14 LeftMFG LeftMidPostTemp
## 16 LeftAngG LeftPostTemp
## 17 LeftIFG LeftPostTemp
## 18 LeftIFGorb LeftPostTemp
## 19 LeftMFG LeftPostTemp
## 21 LeftMidPostTemp LeftPostTemp
## 26 LeftMidAntTemp RightMidAntTemp
## 28 LeftPostTemp RightMidAntTemp
## 29 LeftAngG RightMidPostTemp
## 30 LeftIFG RightMidPostTemp
## 32 LeftMFG RightMidPostTemp
## 35 LeftPostTemp RightMidPostTemp
## 41 LeftMidAntTemp gjt_std_score_gjt_web
## 44 RightMidAntTemp gjt_std_score_gjt_web
## 45 RightMidPostTemp gjt_std_score_gjt_web
## 91 accuracy_children_tsl_accuracies entropy_children_ssl_entropy
## 104 accuracy_children_tsl_accuracies entropy_children_tsl_entropy
## 106 LeftAngG rt_children_ssl_indiv_rts
## 108 LeftIFGorb rt_children_ssl_indiv_rts
## 109 LeftMFG rt_children_ssl_indiv_rts
## 111 LeftMidPostTemp rt_children_ssl_indiv_rts
## 126 LeftMidPostTemp rt_children_tsl_indiv_rts
## 142 LeftMidPostTemp slope_children_ssl_indiv_rts_slope
## 153 rt_children_tsl_indiv_rts slope_children_ssl_indiv_rts_slope
## 154 LeftAngG slope_children_tsl_indiv_rts_slope
## 155 LeftIFG slope_children_tsl_indiv_rts_slope
## 156 LeftIFGorb slope_children_tsl_indiv_rts_slope
## 157 LeftMFG slope_children_tsl_indiv_rts_slope
## 159 LeftMidPostTemp slope_children_tsl_indiv_rts_slope
## 160 LeftPostTemp slope_children_tsl_indiv_rts_slope
## 169 rt_children_ssl_indiv_rts slope_children_tsl_indiv_rts_slope
## 170 rt_children_tsl_indiv_rts slope_children_tsl_indiv_rts_slope
## 172 LeftAngG age_at_neuro_year
## 185 entropy_children_ssl_entropy age_at_neuro_year
## cor p n
## 1 0.813 0.000 16
## 2 0.849 0.000 16
## 3 0.847 0.000 16
## 4 0.688 0.003 16
## 5 0.858 0.000 16
## 6 0.695 0.003 16
## 9 0.559 0.024 16
## 10 0.714 0.002 16
## 11 0.706 0.002 16
## 12 0.733 0.001 16
## 13 0.639 0.008 16
## 14 0.577 0.019 16
## 16 0.643 0.007 16
## 17 0.833 0.000 16
## 18 0.673 0.004 16
## 19 0.706 0.002 16
## 21 0.655 0.006 16
## 26 0.619 0.011 16
## 28 0.530 0.035 16
## 29 0.515 0.041 16
## 30 0.631 0.009 16
## 32 0.655 0.006 16
## 35 0.613 0.012 16
## 41 -0.938 0.018 5
## 44 -0.917 0.028 5
## 45 -0.964 0.008 5
## 91 -0.732 0.010 11
## 104 -0.730 0.007 12
## 106 -0.597 0.031 13
## 108 -0.607 0.028 13
## 109 -0.573 0.041 13
## 111 -0.718 0.006 13
## 126 0.596 0.041 12
## 142 -0.602 0.030 13
## 153 -0.612 0.045 11
## 154 0.739 0.006 12
## 155 0.739 0.006 12
## 156 0.604 0.037 12
## 157 0.665 0.018 12
## 159 0.876 0.000 12
## 160 0.795 0.002 12
## 169 -0.619 0.042 11
## 170 0.659 0.020 12
## 172 -0.524 0.037 16
## 185 0.732 0.004 13# write.csv(iascl_beh_gray_matter_pearson_all, "iascl_beh_gray_matter_pearson_all.csv")Ignore: Gray Matter-Beh Corr TD Only
Partial Correlations Controling for Age and IQ
partial_corr_data <-
merge(iascl_beh_gray_matter, drive_checklist[,c("Participant.ID", "kbit_matrices_std")],
by.x = "part_id", by.y = "Participant.ID",
all.x = TRUE)
partial_corr_data$kbit_matrices_std <- as.numeric(as.character(partial_corr_data$kbit_matrices_std))
partial_corr_data_ssl <- subset(partial_corr_data, !is.na(rt_children_ssl_indiv_rts))
partial_corr_data_tsl <- subset(partial_corr_data, !is.na(rt_children_tsl_indiv_rts))
library("ppcor")## Loading required package: MASS##
## Attaching package: 'MASS'## The following object is masked from 'package:dplyr':
##
## selectpartial_corr_tsl_result <-
pcor.test(partial_corr_data_tsl$LeftMidPostTemp,
partial_corr_data_tsl$rt_children_tsl_indiv_rts,
partial_corr_data_tsl[,c("age_at_neuro_year","kbit_matrices_std")],method="pearson")
partial_corr_tsl_result <- describe(partial_corr_tsl_result)
partial_corr_ssl_result <-
pcor.test(partial_corr_data_ssl$LeftMidPostTemp,
partial_corr_data_ssl$rt_children_ssl_indiv_rts,
partial_corr_data_ssl[,c("age_at_neuro_year","kbit_matrices_std")],method="pearson")
partial_corr_ssl_result <- describe(partial_corr_ssl_result)
library("psych")##
## Attaching package: 'psych'## The following object is masked from 'package:Hmisc':
##
## describe## The following objects are masked from 'package:ggplot2':
##
## %+%, alphapartial_cor_r <- partial.r(data=partial_corr_data[,-c(1,11, 12)],
x=c("LeftMidPostTemp", "rt_children_ssl_indiv_rts"),
y=c("age_at_neuro_year", "kbit_matrices_std"), method = "pearson", use="pairwise")
cp <- corr.p(partial_cor_r,n=11)
print(cp, short = F)## Call:corr.p(r = partial_cor_r, n = 11)
## Correlation matrix
## partial correlations
## LeftMidPostTemp rt_children_ssl_indiv_rts
## LeftMidPostTemp 1.00 -0.71
## rt_children_ssl_indiv_rts -0.71 1.00
## Sample Size
## [1] 11
## Probability values (Entries above the diagonal are adjusted for multiple tests.)
## partial correlations
## LeftMidPostTemp rt_children_ssl_indiv_rts
## LeftMidPostTemp 0.00 0.01
## rt_children_ssl_indiv_rts 0.01 0.00
##
## Confidence intervals based upon normal theory. To get bootstrapped values, try cor.ci
## lower r upper p
## LfMPT-r____ -0.92 -0.71 -0.2 0.01partial_cor_r_tsl <- partial.r(data=partial_corr_data[,-c(1,11, 12)],
x=c("LeftMidPostTemp", "rt_children_tsl_indiv_rts"),
y=c("age_at_neuro_year", "kbit_matrices_std"), method = "pearson", use="pairwise")
cp_tsl <- corr.p(partial_cor_r_tsl,n=10)
print(cp_tsl, short = F)## Call:corr.p(r = partial_cor_r_tsl, n = 10)
## Correlation matrix
## partial correlations
## LeftMidPostTemp rt_children_tsl_indiv_rts
## LeftMidPostTemp 1.00 0.74
## rt_children_tsl_indiv_rts 0.74 1.00
## Sample Size
## [1] 10
## Probability values (Entries above the diagonal are adjusted for multiple tests.)
## partial correlations
## LeftMidPostTemp rt_children_tsl_indiv_rts
## LeftMidPostTemp 0.00 0.01
## rt_children_tsl_indiv_rts 0.01 0.00
##
## Confidence intervals based upon normal theory. To get bootstrapped values, try cor.ci
## lower r upper p
## LfMPT-r____ 0.21 0.74 0.93 0.01Ignore: Gray Matter-Beh Corr ASD Only
Avg Cortical Thickness-Beh Corr (ASD + TD)
iascl_beh_cor_thick_pearson_all <- rcorr(as.matrix(
iascl_beh_cor_thick[,c(2:10, 13:23)]), type = c("pearson"))
iascl_beh_cor_thick_pearson_all <-
flattenCorrMatrix(iascl_beh_cor_thick_pearson_all$r,
iascl_beh_cor_thick_pearson_all$P,
iascl_beh_cor_thick_pearson_all$n)
iascl_beh_cor_thick_pearson_all[,c(3:5)] <-
round(iascl_beh_cor_thick_pearson_all[,c(3:5)], 3)
iascl_beh_cor_thick_pearson_all <-
iascl_beh_cor_thick_pearson_all[which(iascl_beh_cor_thick_pearson_all$p <= 0.05),]
print(iascl_beh_cor_thick_pearson_all)## row column
## 1 LeftAngG LeftIFG
## 3 LeftIFG LeftIFGorb
## 4 LeftAngG LeftMFG
## 5 LeftIFG LeftMFG
## 12 LeftIFG LeftMidPostTemp
## 14 LeftMFG LeftMidPostTemp
## 15 LeftMidAntTemp LeftMidPostTemp
## 16 LeftAngG LeftPostTemp
## 17 LeftIFG LeftPostTemp
## 19 LeftMFG LeftPostTemp
## 29 LeftAngG RightMidPostTemp
## 30 LeftIFG RightMidPostTemp
## 31 LeftIFGorb RightMidPostTemp
## 91 accuracy_children_tsl_accuracies entropy_children_ssl_entropy
## 104 accuracy_children_tsl_accuracies entropy_children_tsl_entropy
## 111 LeftMidPostTemp rt_children_ssl_indiv_rts
## 153 rt_children_tsl_indiv_rts slope_children_ssl_indiv_rts_slope
## 169 rt_children_ssl_indiv_rts slope_children_tsl_indiv_rts_slope
## 170 rt_children_tsl_indiv_rts slope_children_tsl_indiv_rts_slope
## 185 entropy_children_ssl_entropy age_at_neuro_year
## cor p n
## 1 0.499 0.049 16
## 3 0.659 0.006 16
## 4 0.635 0.008 16
## 5 0.772 0.000 16
## 12 0.754 0.001 16
## 14 0.683 0.004 16
## 15 0.581 0.018 16
## 16 0.552 0.027 16
## 17 0.587 0.017 16
## 19 0.733 0.001 16
## 29 0.518 0.040 16
## 30 0.522 0.038 16
## 31 0.532 0.034 16
## 91 -0.732 0.010 11
## 104 -0.730 0.007 12
## 111 -0.633 0.020 13
## 153 -0.612 0.045 11
## 169 -0.619 0.042 11
## 170 0.659 0.020 12
## 185 0.732 0.004 13# write.csv(iascl_beh_cor_thick_pearson_all, "iascl_beh_cor_thick_pearson_all.csv")Ignore: Avg Cortical Thickness-Beh Corr TD Only
Ignore: Avg Cortical Thickness-Beh Corr ASD Only
MRI Gray Matter Corr Plots
SSL Mean RT: Negative Corr between SSL Mean RT and Gray Matter Volume
ggplot(data = subset(iascl_beh_gray_matter, !is.na(rt_children_ssl_indiv_rts)),
aes(x=rt_children_ssl_indiv_rts,
y = LeftAngG,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
ggplot(data = subset(iascl_beh_gray_matter, !is.na(rt_children_ssl_indiv_rts)),
aes(x=rt_children_ssl_indiv_rts,
y = LeftMidPostTemp,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
iascl_beh_mri_corr_plot <-
iascl_beh_gray_matter[, c("part_id",
"LeftMidPostTemp",
"slope_children_ssl_indiv_rts_slope",
"slope_children_tsl_indiv_rts_slope",
"group")]
iascl_beh_mri_corr_plot <- melt(iascl_beh_mri_corr_plot, id.vars=c("part_id", "group", "LeftMidPostTemp"))
iascl_beh_mri_corr_plot[which(iascl_beh_mri_corr_plot$part_id == "blast_c_215"),
"group"] <- "TD"
iascl_beh_mri_corr_plot$variable <-
revalue(iascl_beh_mri_corr_plot$variable, c(#"accuracy_children_lsl_accuracies"="Letter",
#"accuracy_children_vsl_accuracies"="Image",
"slope_children_tsl_indiv_rts_slope"="Non-linguistic",
"slope_children_ssl_indiv_rts_slope"="Linguistic"))
colnames(iascl_beh_mri_corr_plot)[colnames(iascl_beh_mri_corr_plot) == "variable"] <- "Task"
colnames(iascl_beh_mri_corr_plot)[colnames(iascl_beh_mri_corr_plot) == "group"] <- "Group"
iascl_beh_mri_corr_plot$legend <- paste(iascl_beh_mri_corr_plot$Group, iascl_beh_mri_corr_plot$Task)
library("stringr")
iascl_beh_mri_corr_plot$unique_id <- str_extract(iascl_beh_mri_corr_plot$part_id, "(?<=blast_c_)\\S+")
# write.csv(iascl_beh_mri_corr_plot, "iascl_beh_mri_corr_plot_data.csv")
library(ggpubr)## Loading required package: magrittr##
## Attaching package: 'magrittr'## The following object is masked from 'package:pastecs':
##
## extract##
## Attaching package: 'ggpubr'## The following object is masked from 'package:plyr':
##
## mutatelibrary(plyr)
iascl_beh_mri_corr_plot %>%
ggplot(
aes(x=value,
y = LeftMidPostTemp,
color = legend,
shape = legend
)) +
theme(legend.title = element_blank()) +
geom_point(size = 4) +
scale_shape_manual(values=c(16, 16, 17, 17))+
scale_color_manual(values = c("#756bb1", "#fec44f", "#756bb1", "#fec44f")) +
labs(title = "Grey Matter Volume Relationship with Behavioral Task \n",
y = bquote(bold("Left Posterior Middle Temporal Lobe GMV"~(mm^3))), # Change x-axis label
x = "Reaction Time Slope (arbitrary unit/trial)") +
theme(plot.title = element_text(hjust = 0.23)) +
theme(
plot.title = element_text(size=16, face="bold"),
axis.title.x = element_text(size=14, face="bold"),
axis.title.y = element_text(size=14, face="bold"),
axis.text=element_text(size=12, face = "bold")
) +
theme(legend.text=element_text(size=14, face="bold")) +
theme(
panel.background = element_rect(fill = "white"), # Set plot background to white
legend.key = element_rect(fill = "white"), # Set legend item backgrounds to white
axis.line.x = element_line(colour = "black", size = 1), # Add line to x axis
axis.line.y = element_line(colour = "black", size = 1) # Add line to y axis
) +
geom_smooth(data = subset(iascl_beh_mri_corr_plot, Task == "Linguistic"),
aes(x=value, y=LeftMidPostTemp),
method=lm, se=FALSE, show.legend = F, inherit.aes = F, color = "#756bb1") +
geom_smooth(data = subset(iascl_beh_mri_corr_plot, Task == "Non-linguistic"),
aes(x=value, y=LeftMidPostTemp),
method=lm, se=FALSE, show.legend = F, inherit.aes = F, color = "#fec44f") +
stat_cor(data = subset(iascl_beh_mri_corr_plot, Task == "Linguistic"),
aes(x=value, y=LeftMidPostTemp), method = "pearson", label.x = -0.05, label.y = 1590,
inherit.aes = F, color = "#756bb1", size = 5) +
stat_cor(data = subset(iascl_beh_mri_corr_plot, Task == "Non-linguistic"),
aes(x=value, y=LeftMidPostTemp), method = "pearson", label.x = -0.05, label.y = 1550,
inherit.aes = F, color = "#fec44f", size = 5) #+## Warning: Removed 3 rows containing non-finite values (stat_smooth).## Warning: Removed 4 rows containing non-finite values (stat_smooth).## Warning: Removed 3 rows containing non-finite values (stat_cor).## Warning: Removed 4 rows containing non-finite values (stat_cor).## Warning: Removed 7 rows containing missing values (geom_point).
#geom_text(aes(label = unique_id),hjust=0.5, vjust=0.5, color = "black")
# ggsave("iascl_beh_mri_slope_corr_plot.png",
# width = 15, height = 15, units = "cm")MRI Behavioral correlation plot with text labels
library(ggpubr)
library(plyr)
iascl_beh_mri_corr_plot_text <- iascl_beh_mri_corr_plot
iascl_beh_mri_corr_plot_text$legend <-
factor(iascl_beh_mri_corr_plot_text$legend,levels = c("ASD Linguistics",
"TD Linguistics",
"ASD Nonlinguistics",
"TD Nonlinguistics"))
iascl_sl_bar_acc <-
iascl_sl_bar_acc[order(iascl_sl_bar_acc$variable), ]
iascl_beh_mri_corr_plot_text %>%
ggplot(
aes(x=value,
y = LeftMidPostTemp,
color = legend
#shape = legend
)) +
theme(legend.title = element_blank()) +
scale_fill_discrete(labels = c("A", "B", "Linguistics", "Non-Linguistics")) +
geom_point(size = 4) +
scale_shape_manual(values=c(16, 16))+
#scale_color_manual(values = c("#756bb1", "#fec44f", "#756bb1", "#fec44f")) +
#scale_color_manual(values = c("#756bb1", "#fec44f")) +
scale_colour_manual(labels = c("ASD", "ASD", "TD", "TD"),
values = c("#756bb1", "#756bb1", "#fec44f", "#fec44f")) +
# guide = guide_legend(override.aes = list(
# linetype = c(rep("blank", 2), "solid", "dotted"),
# shape = c(rep(16, 2), NA, NA)))) +
labs(title = "Grey Matter Volume Relationship with Behavioral Task \n",
y = bquote(bold("Left Posterior Middle Temporal Lobe GMV"~(mm^3))), # Change x-axis label
x = "Reaction Time Slope (arbitrary unit/trial)") +
theme(plot.title = element_text(hjust = 0.23)) +
theme(
plot.title = element_text(size=16, face="bold"),
axis.title.x = element_text(size=14, face="bold"),
axis.title.y = element_text(size=14, face="bold"),
axis.text=element_text(size=12, face = "bold")
) +
theme(legend.text=element_text(size=14, face="bold")) +
theme(
panel.background = element_rect(fill = "white"), # Set plot background to white
legend.key = element_rect(fill = "white"), # Set legend item backgrounds to white
axis.line.x = element_line(colour = "black", size = 1), # Add line to x axis
axis.line.y = element_line(colour = "black", size = 1) # Add line to y axis
) +
geom_smooth(data = subset(iascl_beh_mri_corr_plot, Task == "Linguistic"),
aes(x=value, y=LeftMidPostTemp), linetype = "solid",
method=lm, se=FALSE, show.legend = T, inherit.aes = F, color = "black") +
geom_smooth(data = subset(iascl_beh_mri_corr_plot, Task == "Non-linguistic"),
aes(x=value, y=LeftMidPostTemp), linetype = "dotted",
method=lm, se=FALSE, show.legend = T, inherit.aes = F, color = "black") +
#stat_cor(data = subset(iascl_beh_mri_corr_plot, Task == "Linguistic"),
# aes(x=value, y=LeftMidPostTemp), method = "pearson", label.x = -0.05, label.y = 1590,
# inherit.aes = F, color = "#756bb1", size = 5) +
#stat_cor(data = subset(iascl_beh_mri_corr_plot, Task == "Non-linguistic"),
# aes(x=value, y=LeftMidPostTemp), method = "pearson", label.x = -0.05, label.y = 1550,
# inherit.aes = F, color = "#fec44f", size = 5) +
geom_text(aes(label = unique_id),hjust=0.8, vjust=0.8, color = "black", face = "bold", size =3) +
geom_text(aes(y = 1590, x = -0.05, label = paste("R =", "0.74*"), fontface = 3),color = "black",size=4) +
geom_text(aes(y = 1550, x = -0.05, label = paste("R =", "-0.71*"), fontface = 3),color = "black",size=4)## Warning: Ignoring unknown parameters: face## Warning: Removed 3 rows containing non-finite values (stat_smooth).## Warning: Removed 4 rows containing non-finite values (stat_smooth).## Warning: Removed 32 rows containing missing values (geom_point).## Warning: Removed 7 rows containing missing values (geom_text).
# ggsave("iascl_beh_mri_slope_corr_plot_with_text.png",
# width = 15, height = 15, units = "cm")TSL Mean RT: Positive Corr between TSL Mean RT and Gray Matter Volume
ggplot(data = subset(iascl_beh_gray_matter, !is.na(rt_children_tsl_indiv_rts)),
aes(x=rt_children_tsl_indiv_rts,
y = LeftMidPostTemp,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
TSL Slope: Positive Corr between TSL RT Slope and Gray Matter Volume
ggplot(data = subset(iascl_beh_gray_matter, !is.na(slope_children_tsl_indiv_rts_slope)),
aes(x=slope_children_tsl_indiv_rts_slope,
y = LeftAngG,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
ggplot(data = subset(iascl_beh_gray_matter, !is.na(slope_children_tsl_indiv_rts_slope)),
aes(x=slope_children_tsl_indiv_rts_slope,
y = LeftIFG,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
ggplot(data = subset(iascl_beh_gray_matter, !is.na(slope_children_tsl_indiv_rts_slope)),
aes(x=slope_children_tsl_indiv_rts_slope,
y = LeftIFGorb,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
ggplot(data = subset(iascl_beh_gray_matter, !is.na(slope_children_tsl_indiv_rts_slope)),
aes(x=slope_children_tsl_indiv_rts_slope,
y = LeftMFG,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
ggplot(data = subset(iascl_beh_gray_matter, !is.na(slope_children_tsl_indiv_rts_slope)),
aes(x=slope_children_tsl_indiv_rts_slope,
y = LeftMidPostTemp,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
ggplot(data = subset(iascl_beh_gray_matter, !is.na(slope_children_tsl_indiv_rts_slope)),
aes(x=slope_children_tsl_indiv_rts_slope,
y = LeftPostTemp,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
MRI Cortical Thickness Corr Plots
SSL Mean RT: Negative Corr between SSL Mean RT and Cortical Thickness
ggplot(data = subset(iascl_beh_cor_thick, !is.na(rt_children_ssl_indiv_rts)),
aes(x=rt_children_ssl_indiv_rts,
y = LeftMidPostTemp,
color = group)) +
geom_point() +
scale_fill_brewer(palette = "Set2")
Gender MRI
drive_list <- read.csv("/Users/jojohu/Documents/Qlab/iascl/drive_checklist.csv")
drive_list <- drive_list[,c("Participant.ID", "Group", "Gender", "date.of.birth",
"kbit_matrices_raw", "kbit_matrices_std", "date_of_lab_assessment", "date_of_mri" )]
demo_mri <-
merge(iascl_beh_cor_thick, drive_list,
by.x = "part_id", by.y = "Participant.ID",
all.x = TRUE)
demo_mri_long <- melt(demo_mri, id.vars = c("part_id", "group"))## Warning: attributes are not identical across measure variables; they will
## be droppeddemo_mri_gender_long <- demo_mri_long[which(demo_mri_long$variable == "Gender"),]
demo_mri_gender_long <- subset(demo_mri_gender_long, part_id != "blast_c_215")
add_row <- data.frame(matrix(ncol = 4, nrow = 1))
colnames(add_row) <- colnames(demo_mri_gender_long)
add_row$part_id <- "blast_c_215"
add_row$group <- "TD"
add_row$variable <- "Gender"
add_row$value <- "F"
demo_mri_gender_long <- rbind(demo_mri_gender_long, add_row)
gender_mri_wide <- cast(demo_mri_gender_long, group~value, length)
print(gender_mri_wide)## group F M
## 1 ASD 1 3
## 2 TD 9 3print(chisq.test(gender_mri_wide))## Warning in chisq.test(gender_mri_wide): Chi-squared approximation may be
## incorrect##
## Pearson's Chi-squared test with Yates' continuity correction
##
## data: gender_mri_wide
## X-squared = 1.4222, df = 1, p-value = 0.233Gender is matched in the MRI pariticipants.
KBITT MRI
demo_mri_kbitt_long <- demo_mri_long[which(
demo_mri_long$variable == "kbit_matrices_raw" |
demo_mri_long$variable == "kbit_matrices_std"),]
demo_mri_kbitt_long$value <- as.numeric(as.character(demo_mri_kbitt_long$value))
demo_mri_kbitt_long[which(demo_mri_kbitt_long$part_id == "blast_c_215"), "group"] <- "TD"
library("reshape")
kbitt_mri_group_mean <- cast(demo_mri_kbitt_long, group~variable, "mean")
print(kbitt_mri_group_mean)## group kbit_matrices_raw kbit_matrices_std
## 1 ASD 30.50000 112.7500
## 2 TD 31.58333 111.3333To Do:
-demo info for MRI - Age for in lab MRI appending. -Ancova for gender match - Done -Ancova for Whole brain? Still? No corr between etiv and CT and only one region significant in GM. -make the t-test results into a nice format please